DEPRESSION AND ANXIETY 33:575–583 (2016)
Research Article SLEEP AND TREATMENT OUTCOME IN POSTTRAUMATIC STRESS DISORDER: RESULTS FROM AN EFFECTIVENESS STUDY Miriam J. J. Lommen, Ph.D.,1,2 ∗ Nick Grey, DClinPsy.,3 David M. Clark, DPhil.,1,2,3 Jennifer Wild, DClinPsy.,1,2 Richard Stott, DClinPsy.,3 and Anke Ehlers, Ph.D.1,2,3
Background: Most patients with posttraumatic stress disorder (PTSD) suffer from sleep problems. Concerns have been raised about possible detrimental effects of sleep problems on the efficacy of psychological treatments for PTSD. In this study, we investigated the relation of session-to-session changes in PTSD symptoms and sleep, and tested whether sleep problems predicted poorer short- and long-term treatment outcome. Methods: Self-reported sleep quality, sleep duration, and PTSD symptoms were assessed weekly in a consecutive sample of 246 patients who received cognitive therapy for PTSD (CT-PTSD; Ehlers & Clark, 2000), and at follow-up (mean = 247 days posttreatment). Additionally, moderating effects of medication use and comorbid depression were assessed. Results: Sleep and PTSD symptoms improved in parallel. The relation was moderated by depression: Sleep problems at the start of therapy did not predict improvement in PTSD symptoms during treatment for patients without comorbid depression. Patients with comorbid depression, however, showed less rapid decreases in PTSD symptoms, but comparable overall outcome, if their sleep quality was poor. Residual sleep problems at the end of treatment did not predict PTSD symptoms at follow-up once residual PTSD symptoms were taken into account. Conclusions: CT-PTSD leads to simultaneous improvement in sleep and PTSD symptoms. Sleep problems may reduce the speed of recovery in PTSD patients with comorbid depression. For these patients, additional treatment sessions are indicated to achieve comparable outcomes, and additional interventions targeting sleep may be beneficial. For those without comorbid depression, self-reported sleep problems did not interfere with response to trauma-focused psychological treatment. C 2015 The Authors. Depression and Depression and Anxiety 33:575–583, 2016. Anxiety published by Wiley Periodicals, Inc.
Key words: PTSD/posttraumatic stress disorder; treatment; sleep disorders; depression; CBT/cognitive behavior therapy
A
1 Department
of Experimental Psychology, University of Oxford, Oxford, UK 2 Oxford NIHR Cognitive Health Clinical Research Facility, Oxford, UK 3 National Institute for Health Research (NIHR) Mental Health Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London, London, UK Contract grant sponsor: Wellcome Trust; Contract grant number: 069777.
INTRODUCTION
fter experiencing a traumatic event such as disaster, assault, severe accidents, or war zone experiences, many ∗ Correspondence
to: Miriam J. J. Lommen, Department of Experimental Psychology, University of Oxford, 9 South Parks Road, Oxford OX1 3UD, United Kingdom. E-mail: [email protected] Received for publication 30 March 2015; Revised 5 August 2015; Accepted 16 August 2015 DOI 10.1002/da.22420 Published online 22 September 2015 in Wiley Online Library (wileyonlinelibrary.com).
C 2015 The Authors. Depression and Anxiety published by Wiley Periodicals, Inc. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
576
Lommen et al.
people will have problems sleeping. Among those who develop posttraumatic stress disorder (PTSD), 70–91% also have longstanding sleep disturbances, as evidenced by increased self-reported sleep problems, and objective indicators of poorer sleep including less slow wave sleep and increased stage 1 sleep and rapid eye movement density.[1] Sleep disturbances are seen as a core symptom of PTSD,[2] and constitute one of the diagnostic criteria for PTSD according to the Diagnostic and Statistical Manual of Mental Disorders.[3, 4] An emerging view in the literature, however, suggests that sleep disturbances may not just be a symptom of PTSD, but a correlated independent problem that may interfere with the efficacy of PTSD treatments and needs to be addressed directly in treatment to maximize clinical outcomes.[5] Some experts argue that insomnia is a mechanism involved in the development and maintenance of PTSD.[5] This view is supported by longitudinal studies showing that sleep problems before and shortly after trauma predict PTSD,[6–11] clinical studies showing that pretreatment self-reported sleep disturbances predict lower remittance rates in primary care patients with PTSD[12] and that treatment of insomnia leads to improvement in PTSD symptoms.[13, 14] The involvement of sleep in emotional and memory processing[15] has raised concerns about detrimental effects of sleep disturbances on cognitive behavioral therapy (CBT) efficacy.[16] Experimental studies indicate that sleep promotes retention of fear extinction and extinction generalization,[17–19] although not all studies have found an effect of sleep on extinction memory.[20] Most trauma-focused psychological PTSD treatments include some elements of exposure to trauma memories and trauma reminders. As extinction learning is seen as a mechanism of exposure therapy,[21] an important clinical question is whether sleep disturbances impair therapeutic gains of trauma-focused psychological PTSD treatments. There is as yet insufficient evidence on the relation between sleep and PTSD, and the effects of sleep problems on treatment effects in people with PTSD. If sleep disturbance is a symptom of PTSD, sleep would be expected to improve concurrently with other PTSD symptoms during PTSD treatment. CBT for PTSD has indeed a moderate effect on sleep disturbance according to a meta-analysis (ES = 0.404).[22] Insomnia decreases in PTSD patients who show a clinical significant response to PTSD treatment,[23, 24] and improvements may be maintained for years after treatment completion.[25] However, sleep disturbance, in particular insomnia, does not seem to benefit as much as other PTSD symptoms from empirically supported PTSD treatments,[5, 26] and insomnia is the most common residual symptom that typically remains in the clinical range when PTSD is targeted in treatment.[23, 24] Moreover, there is a growing body of literature showing that insomnia is predictive of other psychiatric disorders including new-onset and recurrent depression.[27–29] Comorbid depression is common in Depression and Anxiety
PTSD (typically around 50%).[30] There is a large literature linking sleep problems and depression.[29] It is thus of clinical importance that insomnia is effectively targeted in the treatment of PTSD with and without comorbid depression. Previous studies on the relation between changes in sleep and PTSD during therapy have been limited by focusing on pre- to posttreatment changes, consequently prohibiting any analyses on the course of symptoms across therapy. To investigate the nature of the relation between sleep and PTSD more closely, this study analyzed session-to-session changes in sleep duration and quality and PTSD symptoms in a consecutive sample of patients who received cognitive therapy for PTSD (CTPTSD).[31] The aims of this study were to investigate (1) the relation between changes in PTSD symptoms and sleep across treatment, (2) whether sleep problems at the beginning of treatment predicted poorer response to treatment, and (3) whether residual sleep problems at the end of treatment predicted poorer long-term outcome, over and above what could be predicted from PTSD symptoms at the end of treatment. Expanding on research question 2, this study also explored the effects of medication use and comorbid depression, as both might affect sleep and PTSD pathology. Findings with regard to the effect of these variables on treatment outcome in patients with PTSD have been mixed, with medication use as a moderator of treatment outcome in some studies,[32] but not others,[33] and comorbid depression as a moderator of treatment outcome in some studies,[34,35] but only a nonspecific predictor in others.[32, 33] This study explored whether medication use or a comorbid major depressive disorder at the start of treatment moderated the effects of sleep problems on PTSD symptom change. Investigation of these research questions can inform therapists whether it is necessary to specifically target sleep problems in patients with PTSD.
METHOD PARTICIPANTS AND PROCEDURE The patients eligible for this study constitute a consecutive subsample (n = 246) of a cohort study of 330 patients aged 17–83 who received psychological treatment for PTSD at a National Health Service outpatient clinic serving a defined catchment area in South London between April 2001 and August 2008. Weekly sleep measures were introduced from September 2003 onwards. The local research ethics committee approved the study. Details can be found elsewhere.[33] Patients met diagnostic criteria for PTSD according to the Structured Clinical Interview for DSM-IV (SCID).[36] On average, the traumatic event addressed in treatment happened 37.43 months before initial assessment (range 2.50–360 months). At initial clinical interview, 91.9 % of the patients reported sleep problems, with 72.2% reporting sleep onset problems, 91.9% mid sleep awakening, and 57.5% early morning awakening. See Table 1 for demographic information. The majority of the patients included in this study (n = 158, 64.23%) also provided data at long-term follow-up, which on average took place at 247 days after the last therapy session. Compared to patients who provided long-term follow-up data, those who did not re-
Research Article: Sleep and PTSD Treatment
TABLE 1. Demographic information (N = 246) n (valid %) Gender Male Marital status Single Married or cohabiting Divorced or widowed Educational degree None GCSE (school qualification taken at age 16) or equivalent professional education A-level degree (school qualification taken at age 18) University level degree Race White Black Multiracial Trauma type sought treatment for Interpersonal trauma Accidents Event where another person was harmed or died Another kind of traumatic experience Comorbid Axis-I disorder Comorbid anxiety disorder Comorbid depression Taking psychotropic medication (stable dose for at least 2 months before treatment started)
98 (40) 103 (43) 90 (38) 47 (20) 41 (17) 99 (42) 39 (17) 56 (24) 138 (56) 63 (26) 44 (18) 141 (57) 59 (24) 19 (8) 27 (11) 195 (79) 107 (44) 129 (52) 102 (46)
ported more PTSD symptoms and poorer sleep quality and duration at the end of the therapy, all Ps < .01, and were more likely to be divorced or widowed, have a lower educational degree, have a race other than white, and have been diagnosed with another Axis-I disorder, all Ps < .05.
TREATMENT Patients received a course of CT-PTSD[37] ; a trauma-focused psychological treatment that is based on Ehlers and Clark’s cognitive model of PTSD.[31] It aims to change cognitive processes that maintain the perception of current threat in PTSD patients, by focusing on the nature of the trauma memory, problematic appraisals, and behavioral and cognitive strategies that maintain the appraisals and problematic memory features. Details of treatment procedures are found at http://oxcadat.psy.ox.ac.uk/downloads%20and%20links. The treatment does not target sleep directly. Patients received on average 10.30, SD = 5.16, weekly therapy sessions, and optional monthly booster sessions, M = 1.81, SD = 1.75. The analyses of session-tosession changes in this study focus on the first 10 therapy sessions to ensure sufficient valid data per data point.
MEASURES Patients were interviewed by a trained clinician with the full SCID at initial assessment and completed the following self-report questionnaires to assess PTSD symptoms and sleep at initial assessment and before each treatment session. This ensured that for each patient, including dropouts, data at the last treatment session (including any booster sessions) were available. Measures were also collected at 6-month and 1-year follow-ups. Data from the last available followup were used for the analysis (M = 247 days after the last session).
577
PTSD Symptoms (Excluding Sleep). PTSD symptom severity was measured with the Posttraumatic Diagnostic Scale (PDS).[38] The PDS is a 17-item self-report questionnaire assessing PTSD symptoms as specified in the DSM-IV.[3] All items are rated on a 4-point scale (0 = never to 3 = 5 times per week or more / almost always). To avoid content overlap with the sleep measures, a revised sum score of 16 items, excluding the item “Difficulty falling or staying asleep,” was used in the analyses (PDSwos ). Sleep Measures. Sleep quality was assessed by asking patients to rate the quality of their sleep (Overall, how well did you sleep?) during the past week on a scale between 0 (not at all well) and 100 (very well). Sleep duration was assessed by asking patients to report the mean number of hours they slept per night in the past week (In the past week, how much sleep did you have per night?). The sleep duration rating correlated strongly with the sleep duration according to clinical interview at initial assessment, r = .83. Depression. Current comorbid major depression was assessed with the SCID,[36] referring to the last month.
DATA ANALYSIS Parallel process latent growth modeling (LGM) was used (Mplus version 7.11)[39] to analyze the session-to-session changes in PTSD symptoms (excluding the sleep item; PDSwos ) and sleep. Participants who provided data on PTSD symptom severity and sleep for at least three of 10 weekly sessions were included in the analyses, resulting in a sample of 207 and 211 patients to assess the relation between sleep quality and PTSD symptom severity, and the relation between sleep duration and PTSD symptom severity, respectively. Missing (sum) scores on PDSwos or sleep variables in the LGM analyses were estimated using full information maximum likelihood. All analyses were conducted both for the relation between PDSwos and sleep quality, and the relation between PDSwos and sleep duration. The parallel process LGM followed a bottom-up strategy, starting with investigating variance (individual differences) between participants in PDSwos or sleep variables starting level (intercept), and variance in the change in PDSwos or the sleep variables across treatment (slope). For more details on the specific models (see online Appendix A). Model fit was evaluated using fit indices chi square, comparative fit index (CFI), and root-mean-square error of approximation (RMSEA). Chi-square values were used to compare model fit of different models. CFI values of >0.90 and RMSEA values of
Research Article SLEEP AND TREATMENT OUTCOME IN POSTTRAUMATIC STRESS DISORDER: RESULTS FROM AN EFFECTIVENESS STUDY Miriam J. J. Lommen, Ph.D.,1,2 ∗ Nick Grey, DClinPsy.,3 David M. Clark, DPhil.,1,2,3 Jennifer Wild, DClinPsy.,1,2 Richard Stott, DClinPsy.,3 and Anke Ehlers, Ph.D.1,2,3
Background: Most patients with posttraumatic stress disorder (PTSD) suffer from sleep problems. Concerns have been raised about possible detrimental effects of sleep problems on the efficacy of psychological treatments for PTSD. In this study, we investigated the relation of session-to-session changes in PTSD symptoms and sleep, and tested whether sleep problems predicted poorer short- and long-term treatment outcome. Methods: Self-reported sleep quality, sleep duration, and PTSD symptoms were assessed weekly in a consecutive sample of 246 patients who received cognitive therapy for PTSD (CT-PTSD; Ehlers & Clark, 2000), and at follow-up (mean = 247 days posttreatment). Additionally, moderating effects of medication use and comorbid depression were assessed. Results: Sleep and PTSD symptoms improved in parallel. The relation was moderated by depression: Sleep problems at the start of therapy did not predict improvement in PTSD symptoms during treatment for patients without comorbid depression. Patients with comorbid depression, however, showed less rapid decreases in PTSD symptoms, but comparable overall outcome, if their sleep quality was poor. Residual sleep problems at the end of treatment did not predict PTSD symptoms at follow-up once residual PTSD symptoms were taken into account. Conclusions: CT-PTSD leads to simultaneous improvement in sleep and PTSD symptoms. Sleep problems may reduce the speed of recovery in PTSD patients with comorbid depression. For these patients, additional treatment sessions are indicated to achieve comparable outcomes, and additional interventions targeting sleep may be beneficial. For those without comorbid depression, self-reported sleep problems did not interfere with response to trauma-focused psychological treatment. C 2015 The Authors. Depression and Depression and Anxiety 33:575–583, 2016. Anxiety published by Wiley Periodicals, Inc.
Key words: PTSD/posttraumatic stress disorder; treatment; sleep disorders; depression; CBT/cognitive behavior therapy
A
1 Department
of Experimental Psychology, University of Oxford, Oxford, UK 2 Oxford NIHR Cognitive Health Clinical Research Facility, Oxford, UK 3 National Institute for Health Research (NIHR) Mental Health Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London, London, UK Contract grant sponsor: Wellcome Trust; Contract grant number: 069777.
INTRODUCTION
fter experiencing a traumatic event such as disaster, assault, severe accidents, or war zone experiences, many ∗ Correspondence
to: Miriam J. J. Lommen, Department of Experimental Psychology, University of Oxford, 9 South Parks Road, Oxford OX1 3UD, United Kingdom. E-mail: [email protected] Received for publication 30 March 2015; Revised 5 August 2015; Accepted 16 August 2015 DOI 10.1002/da.22420 Published online 22 September 2015 in Wiley Online Library (wileyonlinelibrary.com).
C 2015 The Authors. Depression and Anxiety published by Wiley Periodicals, Inc. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
576
Lommen et al.
people will have problems sleeping. Among those who develop posttraumatic stress disorder (PTSD), 70–91% also have longstanding sleep disturbances, as evidenced by increased self-reported sleep problems, and objective indicators of poorer sleep including less slow wave sleep and increased stage 1 sleep and rapid eye movement density.[1] Sleep disturbances are seen as a core symptom of PTSD,[2] and constitute one of the diagnostic criteria for PTSD according to the Diagnostic and Statistical Manual of Mental Disorders.[3, 4] An emerging view in the literature, however, suggests that sleep disturbances may not just be a symptom of PTSD, but a correlated independent problem that may interfere with the efficacy of PTSD treatments and needs to be addressed directly in treatment to maximize clinical outcomes.[5] Some experts argue that insomnia is a mechanism involved in the development and maintenance of PTSD.[5] This view is supported by longitudinal studies showing that sleep problems before and shortly after trauma predict PTSD,[6–11] clinical studies showing that pretreatment self-reported sleep disturbances predict lower remittance rates in primary care patients with PTSD[12] and that treatment of insomnia leads to improvement in PTSD symptoms.[13, 14] The involvement of sleep in emotional and memory processing[15] has raised concerns about detrimental effects of sleep disturbances on cognitive behavioral therapy (CBT) efficacy.[16] Experimental studies indicate that sleep promotes retention of fear extinction and extinction generalization,[17–19] although not all studies have found an effect of sleep on extinction memory.[20] Most trauma-focused psychological PTSD treatments include some elements of exposure to trauma memories and trauma reminders. As extinction learning is seen as a mechanism of exposure therapy,[21] an important clinical question is whether sleep disturbances impair therapeutic gains of trauma-focused psychological PTSD treatments. There is as yet insufficient evidence on the relation between sleep and PTSD, and the effects of sleep problems on treatment effects in people with PTSD. If sleep disturbance is a symptom of PTSD, sleep would be expected to improve concurrently with other PTSD symptoms during PTSD treatment. CBT for PTSD has indeed a moderate effect on sleep disturbance according to a meta-analysis (ES = 0.404).[22] Insomnia decreases in PTSD patients who show a clinical significant response to PTSD treatment,[23, 24] and improvements may be maintained for years after treatment completion.[25] However, sleep disturbance, in particular insomnia, does not seem to benefit as much as other PTSD symptoms from empirically supported PTSD treatments,[5, 26] and insomnia is the most common residual symptom that typically remains in the clinical range when PTSD is targeted in treatment.[23, 24] Moreover, there is a growing body of literature showing that insomnia is predictive of other psychiatric disorders including new-onset and recurrent depression.[27–29] Comorbid depression is common in Depression and Anxiety
PTSD (typically around 50%).[30] There is a large literature linking sleep problems and depression.[29] It is thus of clinical importance that insomnia is effectively targeted in the treatment of PTSD with and without comorbid depression. Previous studies on the relation between changes in sleep and PTSD during therapy have been limited by focusing on pre- to posttreatment changes, consequently prohibiting any analyses on the course of symptoms across therapy. To investigate the nature of the relation between sleep and PTSD more closely, this study analyzed session-to-session changes in sleep duration and quality and PTSD symptoms in a consecutive sample of patients who received cognitive therapy for PTSD (CTPTSD).[31] The aims of this study were to investigate (1) the relation between changes in PTSD symptoms and sleep across treatment, (2) whether sleep problems at the beginning of treatment predicted poorer response to treatment, and (3) whether residual sleep problems at the end of treatment predicted poorer long-term outcome, over and above what could be predicted from PTSD symptoms at the end of treatment. Expanding on research question 2, this study also explored the effects of medication use and comorbid depression, as both might affect sleep and PTSD pathology. Findings with regard to the effect of these variables on treatment outcome in patients with PTSD have been mixed, with medication use as a moderator of treatment outcome in some studies,[32] but not others,[33] and comorbid depression as a moderator of treatment outcome in some studies,[34,35] but only a nonspecific predictor in others.[32, 33] This study explored whether medication use or a comorbid major depressive disorder at the start of treatment moderated the effects of sleep problems on PTSD symptom change. Investigation of these research questions can inform therapists whether it is necessary to specifically target sleep problems in patients with PTSD.
METHOD PARTICIPANTS AND PROCEDURE The patients eligible for this study constitute a consecutive subsample (n = 246) of a cohort study of 330 patients aged 17–83 who received psychological treatment for PTSD at a National Health Service outpatient clinic serving a defined catchment area in South London between April 2001 and August 2008. Weekly sleep measures were introduced from September 2003 onwards. The local research ethics committee approved the study. Details can be found elsewhere.[33] Patients met diagnostic criteria for PTSD according to the Structured Clinical Interview for DSM-IV (SCID).[36] On average, the traumatic event addressed in treatment happened 37.43 months before initial assessment (range 2.50–360 months). At initial clinical interview, 91.9 % of the patients reported sleep problems, with 72.2% reporting sleep onset problems, 91.9% mid sleep awakening, and 57.5% early morning awakening. See Table 1 for demographic information. The majority of the patients included in this study (n = 158, 64.23%) also provided data at long-term follow-up, which on average took place at 247 days after the last therapy session. Compared to patients who provided long-term follow-up data, those who did not re-
Research Article: Sleep and PTSD Treatment
TABLE 1. Demographic information (N = 246) n (valid %) Gender Male Marital status Single Married or cohabiting Divorced or widowed Educational degree None GCSE (school qualification taken at age 16) or equivalent professional education A-level degree (school qualification taken at age 18) University level degree Race White Black Multiracial Trauma type sought treatment for Interpersonal trauma Accidents Event where another person was harmed or died Another kind of traumatic experience Comorbid Axis-I disorder Comorbid anxiety disorder Comorbid depression Taking psychotropic medication (stable dose for at least 2 months before treatment started)
98 (40) 103 (43) 90 (38) 47 (20) 41 (17) 99 (42) 39 (17) 56 (24) 138 (56) 63 (26) 44 (18) 141 (57) 59 (24) 19 (8) 27 (11) 195 (79) 107 (44) 129 (52) 102 (46)
ported more PTSD symptoms and poorer sleep quality and duration at the end of the therapy, all Ps < .01, and were more likely to be divorced or widowed, have a lower educational degree, have a race other than white, and have been diagnosed with another Axis-I disorder, all Ps < .05.
TREATMENT Patients received a course of CT-PTSD[37] ; a trauma-focused psychological treatment that is based on Ehlers and Clark’s cognitive model of PTSD.[31] It aims to change cognitive processes that maintain the perception of current threat in PTSD patients, by focusing on the nature of the trauma memory, problematic appraisals, and behavioral and cognitive strategies that maintain the appraisals and problematic memory features. Details of treatment procedures are found at http://oxcadat.psy.ox.ac.uk/downloads%20and%20links. The treatment does not target sleep directly. Patients received on average 10.30, SD = 5.16, weekly therapy sessions, and optional monthly booster sessions, M = 1.81, SD = 1.75. The analyses of session-tosession changes in this study focus on the first 10 therapy sessions to ensure sufficient valid data per data point.
MEASURES Patients were interviewed by a trained clinician with the full SCID at initial assessment and completed the following self-report questionnaires to assess PTSD symptoms and sleep at initial assessment and before each treatment session. This ensured that for each patient, including dropouts, data at the last treatment session (including any booster sessions) were available. Measures were also collected at 6-month and 1-year follow-ups. Data from the last available followup were used for the analysis (M = 247 days after the last session).
577
PTSD Symptoms (Excluding Sleep). PTSD symptom severity was measured with the Posttraumatic Diagnostic Scale (PDS).[38] The PDS is a 17-item self-report questionnaire assessing PTSD symptoms as specified in the DSM-IV.[3] All items are rated on a 4-point scale (0 = never to 3 = 5 times per week or more / almost always). To avoid content overlap with the sleep measures, a revised sum score of 16 items, excluding the item “Difficulty falling or staying asleep,” was used in the analyses (PDSwos ). Sleep Measures. Sleep quality was assessed by asking patients to rate the quality of their sleep (Overall, how well did you sleep?) during the past week on a scale between 0 (not at all well) and 100 (very well). Sleep duration was assessed by asking patients to report the mean number of hours they slept per night in the past week (In the past week, how much sleep did you have per night?). The sleep duration rating correlated strongly with the sleep duration according to clinical interview at initial assessment, r = .83. Depression. Current comorbid major depression was assessed with the SCID,[36] referring to the last month.
DATA ANALYSIS Parallel process latent growth modeling (LGM) was used (Mplus version 7.11)[39] to analyze the session-to-session changes in PTSD symptoms (excluding the sleep item; PDSwos ) and sleep. Participants who provided data on PTSD symptom severity and sleep for at least three of 10 weekly sessions were included in the analyses, resulting in a sample of 207 and 211 patients to assess the relation between sleep quality and PTSD symptom severity, and the relation between sleep duration and PTSD symptom severity, respectively. Missing (sum) scores on PDSwos or sleep variables in the LGM analyses were estimated using full information maximum likelihood. All analyses were conducted both for the relation between PDSwos and sleep quality, and the relation between PDSwos and sleep duration. The parallel process LGM followed a bottom-up strategy, starting with investigating variance (individual differences) between participants in PDSwos or sleep variables starting level (intercept), and variance in the change in PDSwos or the sleep variables across treatment (slope). For more details on the specific models (see online Appendix A). Model fit was evaluated using fit indices chi square, comparative fit index (CFI), and root-mean-square error of approximation (RMSEA). Chi-square values were used to compare model fit of different models. CFI values of >0.90 and RMSEA values of
